Last week, President Donald Trump hosted Chinese President Xi Jinping at the White House.
asserted
Trump → host → House
It's their second meeting this year, following the May summit in Beijing.
uncertain
It → follow → Beijing
We have learned about security incidents from nearly every major AI company, American and Chinese, built on both open and closed models.
asserted
We → learn → models
The leading AI labs are openly debating whether to pace their frontier model development, and the government is working out what role it plays in that decision.
asserted
it → lead → decision
I spent four years as Secretary of Commerce focused intently on America’s economic, technological, and security competition with China—instituting first-of-its-kind export controls on the sale of chips and chipmaking equipment to China to protect our national security, and leading implementation of the CHIPS Act to onshore chip manufacturing.
asserted
I → spend → manufacturing
I also created the AI Safety Institute so we could accelerate our AI development responsibly and safely.
uncertain
we → create → development
I have sat across the table from China, and I have an understanding of how they think about AI.
asserted
they → sit → AI
I believe AI companies are right to want to pace the frontier.
asserted
companies → believe → frontier
But we need to stay ahead of China while we pace, because a world where China leads the global AI competition threatens America’s interests.
asserted
China → need → interests
The challenge is this: We need to outcompete China on the technology itself, while also working with China on addressing the risks posed by AI.
asserted
We → need → AI
Competition and cooperation are often in tension, but the stakes are too high in this case to pick only one.
asserted
stakes → pick → one
How we cooperate—while staying ahead
China has its own reasons to take AI safety seriously.
asserted
China → cooperate → safety
The Chinese Minister of State Security recently said publicly that the greatest risk advanced AI poses to China is the risk it poses to the stability of the Chinese Communist Party (CCP).
asserted
it → say → Party
Beijing is no more interested in powerful AI systems wreaking socio-political havoc than Washington is; both countries have incentives to align on certain safety standards or even just a shared understanding of AI risks.
asserted
countries → wreak → risks
It is in neither country’s interest for companies to lose control of their AI systems or for AI to be used to create bioweapons by rogue actors.
asserted
AI → lose → actors
Treasury Secretary Scott Bessent’s announcement of a possible AI dialogue with Chinese counterparts after his meeting with Chinese Vice Premier He Lifeng is a tangible sign that China sees at least some space for cooperation here.
asserted
China → see → cooperation
I would argue that now is the moment to fully establish that dialogue and keep the momentum going.
asserted
momentum → argue → dialogue
We should do this by convening experts from both countries.
asserted
We → do → countries
I’m well aware that we’d be foolish to trust People's Republic of China (PRC) leaders.
asserted
we → ’m → leaders
But that doesn’t mean we shouldn’t try.
asserted
we → mean → ?
It just means we need to develop verification mechanisms to ensure agreements work even in the absence of trust.
asserted
agreements → mean → trust
Even if we don’t reach agreement on a shared set of standards, starting a dialogue and opening those lines of communication are critical—just as we did in 2008 during the financial crisis, when it was in our mutual benefit to coordinate.
asserted
it → reach → benefit
And we spent decades during the Cold War in careful, ongoing dialogue with the Soviet Union about nuclear arms control because the risks to both countries were enormous.
asserted
risks → spend → countries
During that time we developed zero-trust methods to verify arms control agreements even while locked in an adversarial relationship.
asserted
we → develop → relationship
None of that changes the need to vigilantly and continuously update our controls as technology changes so China can’t access our most sophisticated compute in ways that threaten our national security.
asserted
that → change → security
That includes continuously updating our export controls on leading-edge chips, addressing the risks of China accessing controlled compute remotely in data centers around the world, and cracking down on diversion of technology that is controlled.
asserted
that → include → technology
We also need to tighten remaining gaps in our semiconductor manufacturing equipment controls so China can less capably obtain tools to manufacture powerful chips at scale and service their pre-existing equipment.
asserted
China → need → equipment
That work has to happen with our Dutch, Japanese, and German allies and partners who manufacture and service this chipmaking equipment.
asserted
who → have → equipment
Finally, the U.S. government should put a greater burden on the AI companies to take adversarial distillation much more seriously.
asserted
government → put → distillation
Distillation—coordinated, industrial-scale efforts by AI labs to train their weaker models on the outputs of frontier models—has become a significant issue in the past year.
asserted
Distillation → coordinate → year
Anthropic has attributed more than 16 million Claude exchanges to distillation campaigns, and has estimated that over 3 million of these distillation campaigns can be traced back to China’s Moonshot AI (the company behind China’s version of ChatGPT, Kimi) along with other Chinese AI labs like Alibaba and DeepSeek.
asserted
million → attribute → Alibaba
Moonshot, for instance, was later caught forwarding user prompts to Claude to answer and feeding those answers back through Kimi.
asserted
Moonshot → catch → Kimi
China has certainly indigenized much of its AI stack, something Xi publicly made a priority in his Made in China 2025 plan, many years before the U.S. government was talking openly about chips or models.
asserted
government → indigenize → chips
And China has made progress on this plan.
asserted
China → make → plan
But it is clear to many that China’s most competitive models, to some degree, are ripping off American ones.
asserted
models → rip → ones
American AI labs should be allocating even greater amounts of their money and compute towards finding and exposing this kind of activity so that policymakers—and the public—can adapt accordingly.
asserted
policymakers → allocate → activity
The risks and benefits of AI are enormous, something more and more Americans have come to realize as AI advances more quickly than we previously thought possible.
asserted
we → come → AI
There is much to do on AI safety at home, to ensure our society and economy are resilient to the dramatic changes AI will bring.
asserted
AI → be → changes
But I also know that safe deployment of AI is only possible if we successfully manage our relationship with China—both by maintaining our technological lead and working towards collaboration.
asserted
we → know → collaboration
The Aporia
